Meta-analysis: serum vitamin D and breast cancer risk

Abstract

We reviewed and summarised observational epidemiological studies regarding the association between serum vitamin D (measured as 25(OH)D levels) and the risk of breast cancer (BC). Relevant studies published until September 2009 were identified by systematically electronic searching Ovid Medline, EMBASE and ISI Web of Knowledge databases and by cross-referencing. The following data were extracted in a standardised manner from eligible studies: first author, publication year, country, study design, characteristics of the study population, duration of follow-up, BC incidence/BC mortality according to serum 25-hydroxyvitamin D (25(OH)D) and the respective ratios, and covariates adjusted for in the analysis. All existing observational epidemiological studies that reported at least one serum 25(OH)D level in subjects in any time period before or after a diagnosis of breast cancer were included in our review. Individual and summary risk ratios (RRs) for an increase of serum 25(OH)D by 20ng/ml were calculated using meta-analysis methods. Only 25(OH)D was considered. Overall, 10 articles were included. Specific results for BC incidence were reported in nine articles and for BC mortality in one article. In meta-analyses, summary RRs (95% confidence interval (CI)) for an increase of 25(OH)D by 20ng/ml were 0.59 (0.48-0.73), 0.92 (0.82-1.04) and 0.73 (0.60-0.88) with P values of <0.001, 0.164 and 0.001 for case-control studies, nested case-control studies and both study designs combined, respectively. No indication for publication bias was found, but there was large heterogeneity between studies. In conclusion, while case-control studies with measurement of 25(OH)D after diagnosis suggest an inverse association, a statistically significant inverse association remained unconfirmed in prospective studies with measurement of 25(OH)D years before diagnosis. Further studies are needed to clarify the potential role and the relevant exposure time regarding vitamin D and breast cancer risk.